When removing the free wheel diodes, the built-in source to drain diodes of MOSFETs act as a free wheel, because these diodes show a specified avalanching effect, which allow a very fast decaying of motor current. This, because the avalanching effect takes place at an Uds of about 100V and because the higher the inductive kick is allowed to develop, when free wheeling, the faster the motor current decays!

Instead of the standard free wheel diode configuration, you can use a different methode to get a speeded-up decaying of motor current, which is discussed here

http://www.st.com/stonline/...s/1378.pdf (figure 19) and here

http://www.st.com/stonline/...n/1660.pdf (figure 2), for instance.

This methode uses a zener in series with the free wheel diodes, which allows a higher inductive kick and by this a faster decaying of motor current. For the zener you can use an unidirectional transzorb of course.


I see two possible reasons for it:

1. Because the free wheel current is shunted to ground by the internal and avalanching source to drain diodes,

2. and/or because the inductive kick allowed by the internal and avalanching source to drain diodes is too high and noisy arcings take place.

So, you should use a scheme, which (a) shunts the inductive kick not to ground but back to the motor terminals and (b) you should take a transzorb, which limits the inductive kick to an Uds of considerably less than 100V.

Or maybe it's just the scheme showed here

http://www.st.com/stonline/...s/1378.pdf (figure 20, with a turn-on voltage of transzorb of distinctly less than 100V)

which yields an optimum compromise between torque and noise?

Check it out...

Kai